Analysis of LIDAR-Derived Bare Ground Model Accuracy in Southern California Chaparral

LIDAR is developing as a useful tool for measuring biomass. Accurate measurements of biomass may be useful in predicting both the behavior and intensity of wildfire. A better understanding of wildfire behavior in the fire-prone and densely populated areas around Los Angeles and San Diego would be of great benefit. The chaparral ecosystem of Southern California provides a unique operational setting for LIDAR systems in which assumptions of laser behavior may not hold true, calling into question the accuracy of LIDAR-based biomass estimates. A study was undertaken to determine the accuracy of LIDAR-derived bare ground models as a precursor to using LIDAR above-ground returns for biomass estimation. A real-time kinematic GPS survey of five research sites in the San Bernardino and Los Padres National Forests with vegetation cover ranging from low shrubs with grass to dense, mature chaparral was undertaken in June 2007. Measurements of ground elevation, vegetation height, and vegetation density were taken at 1709 points, and these measurements were compared to LIDAR data collected in August of 2005. Methods were used to account for vegetation growth and systematic differences between the LIDAR and GPS data. Overall it was shown that the LIDAR ground elevations were 0.04 ± 0.61 meters above the GPS ground elevations. In the highest density vegetation, the LIDAR ground elevations were 0.10 ± 0.69 meters above the GPS ground elevations, and this increased to 0.13 ± 0.70 meters in the tallest vegetation category. This suggests that laser pulses are often not reaching the ground in this low dense vegetation, but get close. The scale and texture of the chaparral have an effect on LIDAR’s ability to model the ground. In homogenous chaparral with few canopy gaps, LIDAR ground elevations were 0.43 ± 0.78 meters above the GPS ground elevations.